Results?

For about a year, nothing much has happened in this group. Is anything yet to be expected by way of results? Or the whole matter to be considered shelved?

Yeah this thing really fizzled out over the winter and spring. It certainly looks dead in the water.

If Tim hasn't given up on the idea, I still think it would be worth rebooting and starting over with new leadership and a new model.

I think the current project made a dead-end for itself by adapting the BISAC model. Instead of OSC, it could be called BISAC, Only Complicated (BOC.) There's no compelling reason why a library would want to switch from Dewey to this. It would make more sense for a library to just switch to BISAC itself (or come up with its own modified BISAC) than to follow this crazy monster.

But I think there are some things that could be salvaged in a reboot (namely material and audience facets.) And I still think the original impetus for this project (an open source classification system for the 21st century) is still very valid. It's a shame that we are where we are right now.

Tim, are you still interested in this? I am if you are.

Even if all interest has died down, it should still be useful to have a final statement or report on the findings in this project and the reasons for its eventual abandonment or infeasibility or whatever, if not downright failure.

I think the project needs to be rebooted, only this time with the leaders and the Thingamabrarians agreeing on the basic model. The past one was dead in the water once it became clear that the librarian-leaders were operating under the model of "we create a classification, and may or may not take into account or even acknowledge suggestions, even if there's an overwhelming consensus that we disagree with" -- in short, top-down -- while most participants were operating with a model of "we're all in this together, let's see what we can come up with" -- a bottom-up model. Participants got irritated by the lack of communication from the leaders; they'd swoop in once a month or so to make proclamations that usually showed no indication of even having read the discussions, and the leaders probably got irritated at the participants' unwillingness to just accept their decisions without question. I'm not saying either model is necessarily the right one, but the clash of expectations was a recipe for disaster. (And, speaking personally, if the intent was a top-down model, why all the requests for comments and the big to-do about being open?)

If it is rebooted, the first step should be deciding on some sort of basic parliamentary procedure. I expect a lot of the perceived lack of communication (though certainly not all of it) was a result of there being no set way to make a decision. Eventually after months of discussion on a topic, somebody just had to pick a side and say it wins.

I think part of the problem might have been that the "leaders" were experts. For one thing, they had every reason to think they knew better than a bunch of random members of a book website. For another thing, they were busy professionals without the time to even read the discussions, much less participate. What I'd like to see is a leader who's just someone with a lot of time on his/her hands. Experts would be far more useful to the project if they're in a position to contribute freely, and to tell non-experts why we're wrong without seeming like they're taking over.

I have to say while I have followed since Day One the goings-on of this project, I was hesitant to jump in. By the time I got over my fear of being rejected for my silly ideas, I felt I had missed too much to be able to catch up and contribute.

I agree wholeheartedly that some sort of "wrap-up" document should be put forth. That will enable future instalments not only to avoid some of the mistakes in the creation of the actual classification system, but also to provide suggestions on future frameworks and how to collaborate in a much more fluid manner.

This would enable newcomers to offer suggestions while not feeling grossly uninformed.

5>

I think part of the problem might have been that the "leaders" were experts. For one thing, they had every reason to think they knew better than a bunch of random members of a book website.

They were experts in library and classification areas. Participants were experts in other areas, not necessarily just "random members of a book website". Subject-matter expertise -- things like, oh, say, knowing that birds are animals -- was considered completely worthless by the "leaders". If they aren't going to admit that there are areas where participants might know more than they do, then why bother with an "open" system at all?

I agree with the problems with top-down design and ignoring feedback.

The other problem I saw was that there were just too many classifications at each level. I recommend everybody look up the classic paper The Magical Number Seven, Plus or Minus Two.

By that criterion, even the Dewey Decimal System had too many levels, but it was, well, decimal.

7>
Exactly. They didn't necessarily listen to "subject-matter" experts, because they had reason to think they knew better about classification.

#5 should be deciding on some sort of basic parliamentary procedure. I expect a lot of the perceived lack of communication (though certainly not all of it) was a result of there being no set way to make a decision.

That isn't a problem as such. A lot of open source is developed with a benevolent dictator making the decisions.

The main problem I saw was two disjunct processes. The community one where we argued long and hard for various points, and then some hidden one where the leaders came and told us what they had agreed and sketched out the next level for us. Sometimes minor changes from the community were accepted but if the community had agreed that something completely different would be better it was ignored. The leaders were just not part of the community - they did not join in the conversations.

If the project is rebooted then the new leaders have to be active in discussions. http://www.librarything.com/topic/58578 sums up people's frustrations (in a relatively short thread) quite well.

As another example of confusion between leadership and participants, it looks like the leaders have indeed called for new project managers on their blog http://openshelvesclassification.blogspot.com/2009/06/new-project-managers-sough...

This was posted back in June with no mention here--where most interested participants check in. Frustrating.

I'm in favor of just blowing this whole thing up and keeping the few good pieces. Trying to refine the monster that the past leadership created will just lead to countless dead-ends and an ultimately useless and unworkable classification scheme.

Dormant topic, live!

I think the original idea to build an Open Library Classification system is still a good one and should be reattempted (is that a word?).

I still like it, but, man, it's hard to do.

Perhaps I just don't remember, but it doesn't seem to me that the problem with currently used systems, Dewey, LCC, etc. was ever really specifically specified (You like that?) beyond being outdated. I'm sure you (and others) have at least a few specific complaints with those old systems.

After all, the more specific the problem, the easier it is to find the answer. Easier to solve a well-defined problem than a vague not-quite-rightness or a doesn't-really-work-anymoreness.

Just thinking out loud. Take it how you like. :-)

I think that we ran into one of the limitations of crowd sourcing with this one. There are projects that can go a bunch of fundamentally different ways depending on the how you think about them. For instance there's no reason that you can't have an encyclopedia in which the creators of various constructed languages each do a write up of the language they constructed. However, if you want to have a general encyclopedia that relies does not discuss primary research, than those articles can't be in it. So as wikipedia grew up and became the latter, it had to get rid of the people who wanted the former.

I think that there are many different kinds of classification systems and we were still arguing about what kind we wanted, while trying to build the different kinds simultaneously. If you could take a subset of the people here who had a common vision and develop some mechanism for insuring that people who sign-up actually share that vision and aren't trying to introduce something new, I think you could get somewhere. But this project didn't have that, nor would I know how to do that except to hand pick people who had the same vision as me and exclude the rest, which is not crowd sourcing, and is basically how classification systems are currently handled.

14>

This was about Dewey more than LC, and the other problem with Dewey is that beyond the top level it's proprietary.

My thoughts on why this attempt went nowhere are back in #4 and #7.

15>

But this project didn't have that, nor would I know how to do that except to hand pick people who had the same vision as me and exclude the rest

Unfortunately in this case the "vision" that the people brought in to run the show had included being demonstrably wrong on matters of fact.

16: I must say that my personal vision includes "bring in applicable subject experts as needed."

You're more generous than I am in calling the leaders who were brought in "experts in library and classification areas". I had one (much better than average) course on classification and 30 years of practical experience and I certainly did not think that they had the theoretical background to be in charge.

One of the things I regret most about the dissolution of the project, is that, as far as I know, the data collected in the experiment was never analyzed. Even if the classification system we were using for the test wasn't optimal, I think there's a wealth of information in there about the way people think about subjects that could be very valuable if someone had the statistical know-how and a theoretical frame to work from could do some analysis.

>16 lorax: - lorax

I agree completely with your previous posts. I think that, at least generally, the bottom-up approach is the better one.

So, about Dewey. It has two problems so far.

One. It is outdated, that is, (I think I remember reading) it is too U.S.-centric.

Two. Beyond the top level it is proprietary.

The second is, I believe, the more important one. This is not because I live in the U.S. Instead, I believe a free and open classification system could be "easily" adapted to any country or region.

18: There are other problems with Dewey that you haven't listed:

- the base model is not only US-centric but Christian-centric, Indo-European language centric, hetereosexist, mysogynist and racist. The group over-seeing it has been chipping away at these problems since Dewey died, but some of them are intractable (for instance, getting rid of Christian-centrism would mean everyone reclassing every religion book they own.)

- it isn't infinitely expandable. Thus when there are new initiatives in science, like nano-technology that are combinations of two or more former topics (in this case physical and chemical engineering) you have to carve a space out for it in the old numbers; you can't just make a new number. Even when it is expandable, the expansions are silly. All exploration of extraterrestial worlds have to go in 999.9 ...

- the law schedule is pretty meaningless unless you either pick a non-standard option or use numbers too large to put on the books

- the literature schedule is just stupid, assuming that you want all the poetry together and all the plays together and all the essays together, rather than you want all the works by one author together. On top of being stupid, it's extremely hard to use as a cataloger, as is the law schedule.

The biggest problem of all, which applies to all classification systems, is that (at least in the United States) they are not considered to supply enough value for their cost. This means that even though the committee in charge keeps trying to update Dewey, no one reclasses. So the old books are still under the old numbers. The two latest messes are the books on lgbt rights which went from 306 to 305 (or the other way, I don't have access to schedules any more) and general religion books which moved from the 290s to the 200s.

The last is the real stopper on the Open Classification System, as it is very unlikely to be adopted by any library, even if it was completed. It would be too expensive to implement.

>19 aulsmith:

This is exactly what is needed prior to any work being done on a new classification system:

Specific complaints/problems with Dewey and other systems, especially from librarians, etc., so that they may be fixed/avoided.

Also, positive suggestions which might be included/implemented.

Perhaps the one is merely the reciprocal of the other. Still, each should be considered.

As to adoption of the OSC by libraries, of course it will not be done overnight. Perhaps it will be adopted first by some small bookstores and libraries whose stocks and collections are small enough or new enough that conversion would not be a great hassle. After that, well, the most terrible wildfire can be started by the most innocent spark.

aulsmith, if you have a wishlist I would love to hear it. Please, wax eloquent.

-Justin

There's no perfect classification system. It's all trade offs.

The best thing about Dewey is how it goes from the general to the specific by adding numbers, which means that if you encounter a book that you don't have enough background to read, you can chop off the last number, move backwards in the stacks and find a more general book on the same topic. Unfortunately, for this to work you need very long numbers which don't fit on standard library labels. So the vast majority of Dewey libraries chop the numbers off at three or five decimals and get meaningless jumble in certain areas (the 398s being an excellent example.)

LC also goes from broad to specific, but to keep the numbers short, the broad categories are randomly assigned, so, unless you memorize the schedule, you often have to guess where you are (the PNs are a particularly good example of the problem).

Faceted classification systems solve the problem of how to class books on multiple topics, but leave you with the problem of where to put them on the shelves.

One of the problems with the bottom up approach is that everyone thinks they can solve what are obvious problems to them in using the schedules they regularly use, but there's really a lot of theoretical work that very few people are aware of (even I stopped following most the research 20 years ago). Someone is going to have to decide about the obvious trade-offs (length of number, naive vs. specialized use, ease of schedule use vs. size of schedule, shelving vs. searching features, to name a few) before any real, usable work could be done.

>21 aulsmith:

"There's no perfect classification system. It's all trade offs."

This must be remembered. Everyone will not be satisfied with the final result. But that's what is so good about this being free and open. Once it is finished, If you don't like it or if it doesn't work for you, feel free to change it so it works for you.

-----

Does anyone have experience with non-Dewey systems, either librarians or patrons? What do you like and/or dislike about it?

22: It doesn't work that way. The trade-offs are fundamental issues at the core of the design. If you don't like them, the only way to change them is to start over.